Jensen Huang links AI power demand to blockbuster revenue potential as he points to 43 new gigawatts in 2027

NVIDIA CEO Jensen Huang quantified the economics of AI infrastructure demand in remarks reported by Yahoo Finance, arguing that the value of additional AI power is rising sharply. Huang drew a comparison between current and future capacity builds, saying that “two gigawatts to us is like $80 billion,” and describing the shift from earlier periods when the same “gigawatt” scale implied far less revenue potential.

In the reporting, Huang also suggested that the industry is moving toward a new phase of buildout. The same coverage states there are “43 new gigawatts coming in 2027,” positioning the next wave of data center expansion as a major determinant of near-term AI capacity and, by extension, demand for the chips and systems used to run AI workloads.

Huang’s valuation framing includes a historical comparison: the article says the value per gigawatt has increased to more than twice what it was in the past. The reporting contrasts “about $18 billion per gigawatt” earlier with the much higher implied figure now, based on Huang’s “two gigawatts” statement that ties capacity expansion to a large dollar amount.

While the comments are presented as an economic estimate, the coverage does not provide details on the specific accounting model behind “$80 billion,” such as whether it refers to total addressable market, expected revenue, or some other measure. It also does not disclose assumptions about pricing, supply availability, data center utilization, or how quickly new capacity will translate into actual AI compute consumption.

Industry context around Huang’s remarks centers on the practical bottleneck for training and running large AI models: power. Gigawatt-scale growth refers to large increases in the electricity capacity available to data centers, and each increment can constrain how much compute the industry can deploy. NVIDIA, as a leading supplier of accelerators used for AI training and inference, is often viewed by investors as a primary beneficiary when data center operators can afford to build out compute clusters at scale.

Still, the precise linkage between new gigawatts and near-term revenue is not fully spelled out in the coverage. The reporting does not specify which customers or regions correspond to the “43 new gigawatts,” nor does it break down how much of that incremental power is expected to be used specifically for AI versus other workloads. As a result, investors are left to interpret the statement as a directional indicator of rising value per unit of capacity rather than a disclosed forecast.